In some of my previous posts, I have talked about the Pareto’s principle. It is also commonly known as the 80/20 rule or 80/20 principle. It was discovered in 1897 by the Italian economist Vilfredo Pareto. He discovered this while looking at the patterns of wealth and income in England. He found that about 80 percent of wealth was held by about 20 percent of the people. Later on, he found similar patterns in many demographics. The key lesson here is to learn the imbalance.
I recently finished reading the book “The 80/20 Principle” by Richard Koch. The book is not well written. It repeats many things and deviates from the main topic as well. However, some parts of the book are worth reading. It got me thinking more about the principle.
The numbers 80 and 20 are not critical. The idea behind the 80/20 principle is more essential: Not all inputs contribute equally to the outputs. This can be observed in many areas of our life. The actual number can be 70/30 or 99/1 or anything. Since one number is for the inputs and the other number is for the outputs, they may not even add up to 100.
The 80/20 rule is not intuitive. We as humans crave for simple explanations. Our minds usually tend to go for linear relations. That is, each input is equally responsible for the outputs. If we reduce our efforts by half, we expect the results to also reduce by half. But this is typically not the case, since the universe is more complex than a linear relation. It is hence possible to reduce efforts by 80% to only sacrifice 20% reductions in the output.
The 80/20 rule was first observed in financial data. But later on, people saw similar imbalances in many areas. For example, most of the crimes were committed by a few criminals. Most of the profits are generated by a few products or by a few customers. The list goes on. It is not necessary that in each area we would see some imbalance. But given the complexity of the world, it would be safer to not assume linear relations between inputs and outputs.
Here are some areas where I try to apply the 80/20 rule:
My research project:
For one of my projects, I started with a regular plan. I listed down a set of tasks to complete and its expected time. Completing all the tasks could take between 6 and 8 months. Nothing wrong with that. This is so far the intuitive approach. It requires some effort, but I wanted to take it to the next level.
One way to do better is by answering a simple question: If I have to do it in just one month, how would I do it? Obviously, it is not possible to do all the tasks that I planned to do in just a month. So, I have to prune some less important tasks. This forces me to think in a non-linear way. I don’t really have hard time constraints on my projects, so I don’t have to remove any tasks, but this exercise showed me the actual importance of the tasks and obviously some tasks were more critical. Using this, I could reorder the same tasks that would produce results faster.
Thanks to the recent conversation with my wife, I also tried to apply the 80/20 rule to determine what makes me more happy. I found that most of my happiness comes from a very few sources:
- Solving puzzles or making some research contribution.
- Deep mind simulating conversations (usually one to one).
- Learning new stuff.
Many other things like watching netflix or hanging out with a group also makes me happy. But they are not as powerful as the ones listed above. The idea is that I try to maximize the activities that increase my happiness. This doesn’t mean that I completely stop doing other activities. Yet, I should think before starting a movie on netflix: Is there anything else I would rather do that would make me more happy?
Personal and professional relations also follow the Pareto’s law. Very few of them contribute significantly to our happiness or career growth. Being an introvert, I have always limited the number of close friends. I rarely make efforts to build better relations with most people. This allows me to spend a ton of time with the few key people in my life that actually matter. Most people find it hard to understand.
The same goes with professional relations. The common wisdom is to keep trying to make contacts with as many people as we can. I find this ridiculous. Having a deeper relation with a few people is always more advantageous. Many people have helped me with my career. Almost all of them have directly worked with me in some way. The deeper relations take time to build. This is one of the reasons why I find it futile to spend time in the conference parties. I have made better connections by talking to individuals in 1:1 conversations.
Other minor stuff
- I find that I wear only a small fraction of the clothes I own, 80% of the time. The rest are just taking up space in my closet. This is also true for most of the items I own.
- The institute email has a very limited storage. This means that we have to clean up our inbox every now and then. I used 80/20 here last time. I sorted the emails by size and gained back most of the space by deleting just a few emails. Some of my fellow students liked this simple trick and are now using them too.
- I find that only a 20 percent of the stocks I own makes up for the 80 percent of the gains. Unfortunately, as of now, it is hard for me to predict which 20 percent will be beneficial in the future. But I would certainly like to educate myself to get better at making such predictions.
- After a little exploration with cuisines, I have found the few dishes I like the most and usually end up exploiting them. Although, the exploration is not at halt. I do try new things once in a while.
Some work related 80/20 observations
This section is a bit technical. So, feel free to skip it if you don't understand much.
- For a long time, I worked on developing combinatorial optimization solvers. I observed that a few fundamental parts of the solver make most of the difference. The rest of the heuristics are helpful, but not as effective. They do make the solver faster, but at the same time makes the code significantly more complex. I wonder if Gurobi or Cplex can reduce their code size to 20% and still keep 80% of the speed.
- My fellow PhD student Matteo is working on reducing the size of neural networks. In a recent conversation about his project, he mentioned a version of the Pareto’s law. People make complex neural networks for their machine learning tasks, but only a small fraction of the neural network is actually triggered for most of the inputs. Hence, we could reduce the size of the networks by a lot, which would allow us to put them on small devices like our mobile phones.
As mentioned before, 80/20 thinking is not intuitive. It often goes against conventional wisdom. It takes a lot of practice to develop. I am curious to know how you have applied the 80/20 rule in your life (even unknowingly). Let me know below!
Video: Stop drinking cold drinks (in Hindi with English captions)
Quote: “It is far more lucrative and fun to leverage your strengths instead of attempting to fix all the chinks in your armor. The choice is between multiplication of results using strengths or incremental improvement fixing weaknesses that will, at best, become mediocre. Focus on better use of your best weapons instead of constant repair.” ― Timothy Ferriss, The 4-Hour Workweek